Selective in situ potential-assisted SAM formation on multi electrode arrays

11-(ferrocenyl)undecanethiol (Sigma Aldrich, USA) was dissolved in absolute ethanol to a concentration of 1 mM. 11-(ferrocenyl)-carbonyl undecanethiol was synthesized as per literature methods (see Supporting Information). Alkanethiol SAM formation involves the use of a 1:1 solution of 11-(ferrocenyl)undecanethiol (Fc–C₁₁–SH) and 11-(ferrocenyl)-carbonyl undecanethiol (Fc–CO–C₁₁–SH) in 100 mM LiClO₄ (absolute ethanol). Electrochemical cleaning was performed in a 50 mM KClO₄ solution. All other cyclic voltammograms were recorded in 100 mM NaClO₄. The following reagents were purchased and used without further purification: potassium perchlorate (>99%, Sigma Aldrich, USA), lithium perchlorate (>95%, Sigma Aldrich, USA), sodium perchlorate (>98%, Sigma Aldrich, USA) and absolute ethanol (>99.8%, Sigma Aldrich, USA).

Silicon wafers were diced into small (0.5 × 1 cm) pieces and solvent-cleaned with acetone, isopropanol and methanol before sequential thermal evaporation of Ti and Au was performed. A 2 nm Ti adhesion layer was evaporated at a constant rate of 0.9 Å s^–1 followed by a 100 nm thick Au layer at a constant rate of 1 Å s^–1 (pressure <1 × 10⁻⁶ mBar, room temperature). Samples were stored under ambient conditions until needed. To define the electrochemical active area of the exposed gold in solution, a thin layer of Eccobond 286 (Emerson & Cuming, USA) is applied to the base of the gold surface leaving an exposed macroscopic area of ∼1.0 mm².

All samples were electrochemically cleaned prior to each experiment. Samples were cycled between −0.8 and 1.4 V (versus Ag/AgCl (sat. KCl)) in 50 mM KClO₄ at a scan rate of 20 mV s^–1 until a stable voltammogram was obtained. The gold electrode served as the working electrode, a platinum wire (1 mm diameter, Alfa Aesar, USA) as the counter electrode, a standard Ag/AgCl (sat. KCl) reference electrode (BASi, USA) as the reference. Electrochemical measurements were performed using a CHI 1030A (CH Instruments, USA) potentiostat. For all experiments two gold electrodes are recorded simultaneously by using each electrode as a separate working electrode sharing one counter and one reference electrode (Bipotentiostat).

The area of the gold electrode was determined from the magnitude of the gold oxide reductive stripping peak in the cyclic voltamogram performed from −0.8 to 1.4 V. The quantity of surface oxide formed during the anodic excursion is determined by integrating the gold oxide reduction peak in the cathodic scan, Q_red. Assuming a standard charge value of 400 μC cm^–2 for polycrystalline gold [29], the microscopic surface area can be calculated by:

$$\begin{eqnarray}&&A=\displaystyle \frac{{Q}_{{\rm{red}}}}{{Q}_{{\rm{st}}}}\mathrm{.}\end{eqnarray} \tag{ 3 }$$

The geometric surface area is determined for each gold electrode via optical methods. The ratio of the electrochemical surface area to the geometric area yields roughness factor values of 1.1–1.6 for the electrodes used in this study.

The electrochemical area determined using equation (3) is then used to determine the ferrocenylalkythiol coverage by integrating the ferrocene-associated oxidation peak in the CV obtained in 100 mM NaClO₄. The surface coverage of the electroactive ferrocene group is given by:

$$\begin{eqnarray}&&{\rm{\Gamma }}=\displaystyle \frac{Q}{{nFA}}\end{eqnarray} \tag{ 4 }$$

with $Q\,=$ charge obtained by integrating the current peak, n is the number of electrons transferred (n = 1 for the ferrocene couple), F the Faraday constant, and A is the electrochemical surface area of the gold electrode [30]. To determine the fractional surface coverage of the ferrocene-modified alkanethiol on the gold electrode, the theoretical maximal coverage of 4.5 × 10⁻¹⁰ mol cm^–2 (equivalent to 2.7 × 10¹⁴ molecules cm^–2) is divided by the calculated coverage [31]. The theoretical maximal coverage is calculated by assuming a spherical ferrrocene head group adopting a hexagonally close-packed geometry with a diameter of 6.6 Å [32].

The gold surfaces were selectively modified with the two different ferrocene derivatives by varying the applied potential as described in detailed below. Modification was performed in either 1 mM Fc–CO–C₁₁–SH or Fc–C₁₁–SH (1:1) with 100 mM NaClO₄ solutions for 10 min at a constant applied potential. After electrodeposition, the surfaces were rinsed with ethanol and Milli-Q water before a CV was recorded in 100 mM NaClO₄ from 0.2 to 0.8 V (versus Ag/AgCl (sat. KCl)).

The first step toward a successful potential-assisted electrode modification is to determine the ferrocenylalkythiol coverage as a function of the applied potential. Prior to each experiment, the evaporated gold electrodes are electrochemically cleaned before exposing them to the ferrocene derivatize [22]. Measurement of the extent of potential-dependent adsorption involved exposure of the gold electrode to a 1 mM Fc C₁₁–SH/100 mM LiClO₄ solution while applying a series of cathodic potentials (versus Ag/AgCl (sat. KCl)) ranging from −1.4 to −0.9 V, for 10 min at each potential. After each incubation step, the sample was rinsed with ethanol and Milli-Q water before a CV was recorded in 100 mM NaClO₄ at 20 mV s^–1. The resulting ferrocene oxidation peak was integrated to calculate the ferrocene surface coverage (equation (4)). Coverage as a function of applied potential (figure 1(A)) reveals that no apparent ferrocene derivative adsorption occurs at potentials <−0.9 V. Potentials of <−0.9 V (red area) thus serve to maintain a gold electrode free of chemisorbed ferrocenylalkanethiol.

Zoom In Zoom Out Reset image size

On the other hand, application of a potential of 0.3 V (versus Ag/AgCl (sat. KCl)) to a clean gold electrode results in Fc–C₁₁–SH coverage of >50% after 10 min incubation.

E_hold, the potential range over which the ferrocenylalkylthiol SAMs are stable in regard to both desorption and thiol-for-thiol exchange is important. To measure the desorption potential of the ferrocene alkanethiol, the modified surface was exposed to the ferrocenyl solution (1 mM) while changing the potential from −0.3 to −0.8 V. The corresponding ferrocene coverage is shown in figure 1(B). To compare the coverage achieved under different experimental regimes, the coverage is expressed relative to the full coverage case (100%). Starting at −0.3 V and proceeding cathodically, the ferrocene surface coverage decreases indicating that goldthiol bond is destabilized and ferrocenylalkylhiol desorbs from the surface to increasing extent. The holding potential, described as a potential where an existing SAM is stable towards additional alkythiol adsorption or desorption for the time necessary to modify another electrode (5 min) is determined to be >−0.6 V.

This potential-dependent ferrocene coverage establishes the three operational potentials of interest to this study: E_ads, E_des, and E_hold. E_ads (>0.3 V) is the potential at which adsorption of the two ferrocenyl C₁₁-alkanethiols occurs. E_des is the potential at which alkanethiol adsorption onto a clean gold electrode is not measurable (<−0.9 V). E_hold (−0.6 V) is at the potential at which a pre-formed ferrocenyl C₁₁ alkylthiol SAM is held to prevent both further thiol desorption, adsorption and thiol-for-thiol exchange. These potential values provide operational potentials at which one can selectively functionalize two different alkylthiol ferrocene markers onto different but spatially proximal gold electrodes.

These three potential regimes allow one to control the functionalization of two proximal electrodes with two different species. Two different ferrocene derivatives, Fc–CO–C₁₁–SH and Fc–C₁₁–SH demonstrate the effectiveness of the selective functionalization protocol based on the desorption and adsorption. The two probes have the same alkyl chain length (C₁₁) and the same chain terminus (thiol), yet one has an added carbonyl group to the ferrocene moiety. This carbonyl shifts the corresponding redox potential anodically by 250 mV with respect to the Fc–C₁₁–SH [33–36]. The Fc–C₁₁–SH exhibits an oxidation peak at E_p = 0.34 V (versus Ag/AgCl), whereas the Fc–CO–C₁₁–SH exhibits the redox peak at E_p = 0.59 V (versus Ag/AgCl). If a mixed SAM is formed on one electrode, two distinct peaks are observed.

These two different ferrocenes are used to experimentally demonstrate the two-step protocol of selective functionalization of the electrodes shown in figure 2. In the first step, both gold electrodes are simultaneously exposed to the same probe by exposure to 1 mM Fc–CO–C₁₁–SH (100 mM LiClO₄) for 10 min. During the functionalization process, electrode 1 is held at 0.3 V (E_ads), whereas electrode 2 is held at −1.4 V (E_des). After incubation, the electrodes are rinsed with ethanol and Milli-Q water and placed in aqueous 100 mM NaClO₄. A cyclic voltammogram verifies the extent of derivatization of each electrode. Each electrode is rinsed with ethanol at the completion of the functionalization process. In a second step, each electrode is placed in the same solution containing the second adsorbate (2 ml, 1 mM Fc–C₁₁–SH, 100 mM LiClO₄) and held for 10 min at specific potentials. During the functionalization process, electrode 1 is held at −0.6 V (E_hold) while electrode 2 is held at +0.3 V (E_ads) to modify it with the second ferrocene probe.

Zoom In Zoom Out Reset image size

The resulting cyclic voltammogram recorded simultaneously on each gold electrode after the second modification step is shown in figure 3. Electrode 1 (blue) is modified with Fc–CO–C₁₁–SH and Electrode 2 (red) is modified with Fc–C₁₁–SH. Oxidation peaks at E_p = 0.59 V (Fc–CO–C₁₁–SH) and E_p = 0.34 V (Fc–C₁₁–SH) provide for the determination of the respective ferrocene coverage. For the Fc–CO–C₁₁–SH derivitazed electrode 1, the overall surface coverage is 43.4 ± 0.4%, corresponding to an alkanethiol density of 1.951 × 10⁻¹⁰ ± 0.016 × 10⁻¹⁰ mol cm^–2. On the other hand, electrode 2 derivitazed with Fc–C₁₁–SH has a coverage of 32.7 ± 0.3% (density: 1.474 × 10⁻¹⁰ ± 0.015  × 10⁻¹⁰ mol cm^–2).

Zoom In Zoom Out Reset image size

Measurable quantities of Fc–C₁₁ are observed on the Fc–CO–C₁₁-derivatized electrode (figure 4). The latter peak results from the competitive deposition of the undesired alkylferrocene in the presence of desired adsorbate. Quantification of the cross-coverage establishes that electrode 1 (modified with Fc–CO–C₁₁–SH) has a coverage of Fc–C₁₁–SH of 1.1 ± 0.2% (surface density: 4.727 × 10⁻¹² ± 0.748 × 10⁻¹² mol cm^–2) and electrode 2 (modified with Fc–C₁₁–SH) shows a cross-coverage of 2.4 ± 0.2% (surface density: 1.071 × 10⁻¹¹ ± 0.066  × 10⁻¹¹ mol cm^–2) of Fc–CO–C₁₁–SH. The coverage and cross-coverage values associated with three independent experiments are summarized in figure 4. The previously analyzed experiment is shown in C. A and B are carried out using the same protocol. The variance in the coverage likely arises from the polycrystalline nature of the gold electrodes and the relatively short incubation times (10 min) of each step. The effect of roughness has not been correlated with the coverage values. The histogram shows that electrode 1 undergoes high Fc–CO–C₁₁–SH coverage whereas electrode 2 shows high Fc–C₁₁–SH modification. The cross-coverage at each electrode is less than 4% in all experiments.

Zoom In Zoom Out Reset image size

Differentiation has to be made between the two cross-coverage values, as they have different origins. In case C, the cross-coverage value of 2.4% of Fc–CO–C₁₁–SH on electrode 2 results from deposition on a clean electrode held at −1.4 V during the first step. The second cross-coverage value is lesser (with a value of 1.1% of Fc–C₁₁–SH on electrode 1) and results from the second modification step while the partially modified surface is being held at ${E}_{{\rm{hold}}}=-0.6\,{\rm{V}}$. Adsorption on a clean non-functionalized electrode can cause complications if additional surfaces are being modified, as each modification step will increase the degree of undesired adsorption. Due to the in situ electrochemical system, electrochemical cleaning can be performed on a set of non-functionalized arrays after every nth modification. This step will ensure that minimal contamination on new blocked surfaces occurs. On the other hand, undesired adsorption that occurs at E_hold is not likely to increase significantly during further modification steps, as the adsorption kinetics are slower at higher coverage.